Agris on-line Papers in Economics and Informatics

Smart Agriculture with Machine Learning and Internet of Things Techniques: Literature Review and Bibliometric Analysis

Karima El Majdouby Mohamed El Mohadab Karim Abouelmehdi DOI: 10.7160/aol.2026.180301
No 3/2026, September
pp. 3 - 16

ABSTRACT

A booming field known as "smart agriculture" has emerged because of the rapid revolution and transformation of agriculture. It makes use of cutting-edge agricultural technologies, like the internet of things and machine learning (ML) algorithms, to enhance farm management tasks, boost agricultural productivity, and lessen environmental impact by empowering farmers to react swiftly and efficiently to climate change. A key component of the shift to smart farming is the Internet of Things (IoT), which makes it easier for sensors and devices to connect and share data. It also gives farmers access to actionable information and knowledge, enabling them to make well-informed decisions using machine learning algorithms to enhance crop management and boost yields. Numerous tasks, including weed identification, fertilization requirement analysis, irrigation adjustment, and pest and soil management, are made possible by the automation made possible by these technologies. They allow farmers to make targeted interventions by continuously monitoring crop health, which lowers labor costs and the environmental impact. The effects of technological advancements, such as machine learning and the Internet of Things, on the agriculture industry are thoroughly examined in this paper. It draws attention to the useful applications of these instruments, which help farmers better, more automatically, and more precisely manage their resources and output.

KEYWORDS

Smart agriculture, machine learning, internet of things, bibliometric

VIEW MORE PDF (.pdf, 19.26 MB)

Protein Consumption and Socioeconomic Determinants in Indonesia: An Empirical Analysis of the Lowest-Income Households

Mohammad Nafi Izzuddin David Forgenie Lia Rohmatul Maula Sukardi Sukardi Nikmatul Khoiriyah DOI: 10.7160/aol.2026.180302
No 3/2026, September
pp. 17 - 29

ABSTRACT

Adequate protein consumption is essential for improving nutritional outcomes in developing countries, yet empirical evidence on the determinants of protein consumption among the poorest households in Indonesia remains scarce. This study investigates the socioeconomic determinants of protein consumption among the lowest-income households using nationally representative data from the 2023 Survei Sosial Ekonomi Nasional (Susenas), focusing on 67,918 households in the lowest income quintile (Q1). Protein-rich foods were classified into ten categories: eight animal-sourced foods (fish, seafood, beef, mutton, poultry, eggs, milk, and other meats) and two plant-based proteins (tofu and tempeh). A binary probit model was applied to examine the influence of household income, household size, and the relative prices of protein-rich foods on the probability of consumption. Marginal effects were calculated to interpret the change in probability associated with a one-unit change in each explanatory variable. The results reveal contrasting patterns across protein sources. Beef consumption is strongly influenced by all three socioeconomic factors — income, household size, and price — reflecting its status as a luxury protein largely inaccessible to the poorest households. In contrast, eggs emerge as a stable and affordable protein source consumed by nearly 79% of Q1 households, remaining largely insensitive to own-price changes and unaffected by household size. Household income significantly increases the probability of consuming nearly all protein categories, while price sensitivity drives poor households to rely more heavily on tofu and tempeh as affordable alternatives. This study provides novel empirical evidence on the probability-based determinants of protein consumption among Q1 households, offering practical insights for food subsidies, local production support, and nutrition education programs in Indonesia.

KEYWORDS

Protein consumption, low-income households, socioeconomic determinants, household nutrition, food security, Indonesia

VIEW MORE PDF (.pdf, 680.89 KB)

Digital Maturity for Sustainable Livestock Process Management

Oksana Makovoz Sergii Lysenko DOI: 10.7160/aol.2026.180303
No 3/2026, September
pp. 31 - 40

ABSTRACT

Digital transformation in livestock enterprises creates managerial value only when digital tools are embedded in business processes and translated into measurable sustainability effects. The paper develops a digital maturity and portfolio prioritization approach for sustainable livestock process management. The research uses a structurally balanced pilot sample of 20 Ukrainian livestock en-terprises, documentary evidence, expert-documentary scoring, KPI baseline comparison and multi-criteria normalization. The results show uneven maturity across strategy, process automation, data analytics, digital architecture and cyber-resilience. The highest priority is assigned to BI and data warehouse initiatives, followed by ERP-based feed logistics integration and operational control solutions.

KEYWORDS

Business process management, digital maturity, digital transformation, key performance indicators, livestock enterprises, sustainability

VIEW MORE PDF (.pdf, 527.38 KB)

Trends in Household Food Expenditures and Income Elasticity in Slovakia in the Pandemic and Post-pandemic Period

Jozef Palkovič DOI: 10.7160/aol.2026.180304
No 3/2026, September
pp. 41 - 50

ABSTRACT

This paper examines the evolution of food security in Slovakia during the pandemic and post-pandemic years (2020-2023), with a focus on economic access to food. Using household data from the Statistical Office of the Slovak Republic, the study examines changes in income, the share of food expenditures, and income elasticity across regions and demographic groups. The findings show that while household incomes remained relatively stable during the pandemic, the post-pandemic period was marked by rising food prices that outpaced wage growth. As a result, the share of income spent on food increased from 14% in 2020 to 17% in 2023. Data suggests differences between regions: Bratislava recorded the lowest share (12%), whereas Nitra reached 20%. Vulnerability was greatest among retirees, while rural households demonstrated slightly better resilience, likely due to their higher levels of self-sufficiency. Although income elasticity improved overall, disparities widened in 2023 between urban and rural areas. These results underline the need for targeted policies to address affordability and regional inequalities in order to strengthen food security under conditions of economic stress.

KEYWORDS

Food security, food expenditure, income elasticity, regional disparities, demographic vulnerability

VIEW MORE PDF (.pdf, 5.48 MB)

Data Management in Agricultural Enterprises: A Framework for Data-Driven Decision Support

Michal Stočes Jan Jarolímek Irena Benešová Pavel Šimek Zdeňka Žáková Kroupová Ludmila Pánková Jiří Vaněk Eva Kánská Martin Havránek Michael Anderle DOI: 10.7160/aol.2026.180305
No 3/2026, September
pp. 51 - 60

ABSTRACT

The growing digitalisation of agriculture is leading to a rapid increase in the volume of data produced by agricultural enterprises; however, its practical use is often limited by fragmented data sources, varying data quality, insufficient metadata, and inconsistent storage and management practices. The aim of this article is to propose a practical framework for the systematic management of data in agricultural enterprises to support its long-term availability, reuse, and integration into decision-making processes. The proposed approach is based on four interlinked steps: identifying data needs and issues, creating and recording a data inventory, drawing up a Data Management Plan (DMP), and implementing it in practice. The framework links technical aspects of data management – such as data formats, metadata, quality, integrity, security, backup and archiving – with the definition of responsibilities for individual staff members and farm (agricultural enterprise) management. For practical implementation, two basic approaches to data storage are distinguished: file systems and data platforms or repositories. The more advanced platform-based approach enables the centralisation of heterogeneous data sources, the automation of their transfer and processing, and subsequently the integration of the acquired data with analytical tools and decision-support systems. The proposed framework thus represents a path from the isolated collection of operational data to its systematic management and utilisation as a long-term information source for both operational and strategic decision-making within an agricultural enterprise.

KEYWORDS

Agricultural data, Data Management Plan, data quality, metadata, precision agriculture, Institutional Data Repositories, Decision Support sSystems, FAIR data principles

VIEW MORE PDF (.pdf, 3.85 MB)

Agri25: Value Creation Rating Model for Agricultural Enterprises - Comparison of Logistic and Discriminant Analysis in Czech Conditions

Jindřich Špička Zdeňka Náglová DOI: 10.7160/aol.2026.180306
No 3/2026, September
pp. 60 - 77

ABSTRACT

This study develops a sector-specific value creation rating model for Czech agricultural enterprises using Economic Value Added (EVA) as the classification criterion. Based on accounting data from the CRIBIS database covering period 2019-2023, we compare logistic regression and discriminant analysis for distinguishing value-creating from value-destroying firms. Results demonstrate that logistic regression provides superior predictive accuracy, with efficient return on invested capital and adequate working capital relative to long-term financing emerging as the primary determinants of positive EVA. The findings confirm that traditional rating models developed for industrial enterprises exhibit limited applicability in agriculture due to the sector's unique characteristics, including Common Agricultural Policy (CAP) subsidies that suppress bankruptcy rates and create distinct financial dynamics. This research highlights the necessity of sector-adapted analytical tools that shift focus from bankruptcy prediction to value creation assessment, providing practical frameworks for financial institutions, farm managers, and policy-makers evaluating agricultural financial performance and sustainability.

KEYWORDS

Agriculture, creditworthiness, logistic regression, discriminant analysis, economic value added, financial health, Czech farms

VIEW MORE PDF (.pdf, 579.97 KB)

Evaluation of AI Predictability on Crop Yield using XAI - An Experimental Approach

Ramesh Varathan P. Kumaresan DOI: 10.7160/aol.2026.180307
No 3/2026, September
pp. 79 - 101

ABSTRACT

Accurate crop yield prediction is essential for improving agricultural planning and sustainability. This study explores the integration of explainable artificial intelligence (XAI) with machine learning models to enhance transparency in agricultural decision-support systems. A comparative evaluation of multiple regression algorithms - Linear, Ridge, Lasso, Elastic Net, Decision Tree, Random Forest, AdaBoost, XGBoost, CatBoost, LightGBM, Gradient Boosting, and K-Nearest Neighbour Regression (KNR) is performed for crop yield prediction. Categorical attributes are transformed using One-Hot Encoding, while model reliability is ensured through 5-Fold Cross-Validation and hyperparameter optimization using Grid Search. Among the evaluated approaches, KNR demonstrated the best predictive capability with an R² value of 0.9827 and reduced errors (MAE:9.03, RMSE:117.58, MSE:13824.73). To improve interpretability, SHAP, LIME, and ELI5 techniques are employed to explain feature contributions and model decision. The proposed framework provides transparent and reliable insights, supporting data-driven agricultural management and sustainable crop production strategies.

KEYWORDS

XAI, SHAP, LIME, CYP, Agriculture, ML, KNR

VIEW MORE PDF (.pdf, 15.56 MB)

Market Equilibrium Modelling of Rice in South Kalimantan Using the Nonlinear Programming Method: Implications for Agricultural Policy

Nuri Dewi Yanti Yudi Ferrianta DOI: 10.7160/aol.2026.180308
No 3/2026, September
pp. 103 - 121

ABSTRACT

Despite being a top global producer, Indonesia’s rice market equilibrium is increasingly threatened by demographic pressure, land conversion, and a growing reliance on imports. This study develops a novel Computable General Equilibrium (CGE) model using General Algebraic Modelling Language (GAMS) and nonlinear programming to analyse how regional production responds to shifting policies and market dynamics. The research specifically focuses on South Kalimantan, a region that serves as a strategic food buffer for the new National Capital (IKN). The model demonstrates strong empirical reliability, with a price calibration deviation of only 1.8% and an overall production error of just -0.7%. Simulation results reveal that while overproduction creates regional surpluses, it does not necessarily lead to lower consumer prices. Conversely, government-imposed price controls were found to trigger significant supply deficits by disincentivising producer output. These findings confirm the model’s utility as a robust tool for evaluating productivity, infrastructure, and stabilisation policies to ensure national food security.

KEYWORDS

CGE-SAM model, GAMS, rice market equilibrium, South Kalimantan, food security, price control.

VIEW MORE PDF (.pdf, 3.87 MB)

Food Inflation and Poverty in Indonesia: How Regional Disparities Shape Economic Hardship

Misbahol Yaqin Naufal Alfaraby Winandita DOI: 10.7160/aol.2026.180309
No 3/2026, September
pp. 123 - 135

ABSTRACT

We examine the impact of rising food prices on poverty in Indonesia by considering regional characteristics. Using panel data from 34 provinces in Indonesia from 2013 to 2024, we apply the Generalized Method of Moments (GMM) to estimate the relationship between food price inflation and poverty rates. We find that rising food prices significantly increase poverty levels. The results are consistent across models. The impact is more severe in rural areas compared to urban areas, indicating greater vulnerability of rural households to food price fluctuations. Moreover, we find that in Western Indonesia which are a more developed regions experiences a lower poverty impact from rising food prices compared to Eastern Indonesia. These findings highlight the need to stabilize food prices to support inclusive development. Strengthening regional food security, ensuring targeted funding, and implementing effective food subsidies are essential to mitigate the impact of food price inflation on low-income populations

KEYWORDS

Food price inflation, poverty, rural poverty, food security

VIEW MORE PDF (.pdf, 1.56 MB)

Feeding Simulation Model: Data Acquisition to Achieve Climate Neutrality Goals

Mairita Zaķe Andris Lapāns Anda Mežgaile Oskars Java DOI: 10.7160/aol.2026.180310
No 3/2026, September
pp. 137 - 146

ABSTRACT

In line with the European Union’s development plans, preserving a healthy environment for future generations is a key priority. Progress towards climate neutrality requires reliable, coherent, high-quality and interoperable data that capture historical trends and emerging societal dynamics while reflecting ecological, economic and technological systems. This paper addresses data acquisition and integration across five sectors: energy, agriculture, land use, land-use change and forestry, industry and waste management. The proposed framework supports Latvia’s national climate neutrality simulation model with harmonised data. The analysis reveals substantial differences in data accessibility, formats, update regularity, and sectoral methodologies. Critical datasets are often available only in unstructured formats, such as PDFs of infographics, limiting automation in data collection. The developed modular workflow and harmonisation system improve model reliability and support evidence-based long-term climate policy design at national and global levels.

KEYWORDS

Data mapping, data acquisition, climate neutrality, data model, data exchange

VIEW MORE PDF (.pdf, 4.86 MB)